Optimal Dynamic Quantizer Design And Stability Analysis Of Networked Control Systems

Networked systems are a kind of systems that use signals transmitted from communication network. For this kind of systems, control signals are quantized by quantizers before transmission, which can convert actual signals into quantized values in given sets, based on parameters set in advance or calculated in real time. In the past few years, studies of networked quantized system have become more and more important in the field of networked system domain. Actually, as extensive use of quantized devices and algorithms in real networked systems, all digital networked systems are networked quantized systems.Networked quantized systems are considered in this paper, we have designed optimal dynamic quantizers and studied stability of such systems, using Lyapunov stability method, matrix inequality method, stochastic process method, switched and delay system methods.A novel optimal dynamic quantizer with scaling factor is proposed. Quantized error can be compensated using the dynamic quantized structure, and its maximum output error can be minimized. With the help of the scaling factor, a sufficient condition for asymptotic stability is given for the system, in terms of matrix inequalities.Networked quantized systems with Bernoulli packet losses are considered in the paper. The optimal dynamic quantizer is obtained, and a sufficient condition for mean square asymptotic stability is given for the system, in terms of matrix inequalities. Moreover, a state feedback controller design method is also proposed. Finally simulation result is given for the proposed method.Networked quantized systems with time-varying delay is considered in the paper. Its optimal dynamic quantizer and sufficient condition for asymptotic stability are given. Simulation result is finally given for the proposed method.Then networked quantized systems with both packet losses and time-varying delay are analyzed under two kinds of models, and a signal selection strategy that can deal with signal disorder is proposed. The optimal dynamic quantizers are obtained, and sufficient conditions for mean square asymptotic and asymptotic stability are given. Simulation results are given for the proposed method.The predictive control problem of networked quantized systems with time-varying delay and packet losses is analyzed. The quantized predictive compensation strategy is studied to compensate delay and packet loss, and a MIMO optimal dynamic quantizer are obtained. The effectiveness of the proposed method are illustrated through both simulation and experimental results by taking a motor as research object.